[linux-2.6-block.git] / drivers / acpi / numa.c

/*
 *  acpi_numa.c - ACPI NUMA support
 *
 *  Copyright (C) 2002 Takayoshi Kochi <t-kochi@bq.jp.nec.com>
 *
 * ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 *
 *  This program is free software; you can redistribute it and/or modify
 *  it under the terms of the GNU General Public License as published by
 *  the Free Software Foundation; either version 2 of the License, or
 *  (at your option) any later version.
 *
 *  This program is distributed in the hope that it will be useful,
 *  but WITHOUT ANY WARRANTY; without even the implied warranty of
 *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 *  GNU General Public License for more details.
 *
 * ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 *
 */

#define pr_fmt(fmt) "ACPI: " fmt

#include <linux/module.h>
#include <linux/init.h>
#include <linux/kernel.h>
#include <linux/types.h>
#include <linux/errno.h>
#include <linux/acpi.h>
#include <linux/memblock.h>
#include <linux/numa.h>
#include <linux/nodemask.h>
#include <linux/topology.h>

static nodemask_t nodes_found_map = NODE_MASK_NONE;

/* maps to convert between proximity domain and logical node ID */
static int pxm_to_node_map[MAX_PXM_DOMAINS]
			= { [0 ... MAX_PXM_DOMAINS - 1] = NUMA_NO_NODE };
static int node_to_pxm_map[MAX_NUMNODES]
			= { [0 ... MAX_NUMNODES - 1] = PXM_INVAL };

unsigned char acpi_srat_revision __initdata;
int acpi_numa __initdata;

int pxm_to_node(int pxm)
{
	if (pxm < 0)
		return NUMA_NO_NODE;
	return pxm_to_node_map[pxm];
}

int node_to_pxm(int node)
{
	if (node < 0)
		return PXM_INVAL;
	return node_to_pxm_map[node];
}

static void __acpi_map_pxm_to_node(int pxm, int node)
{
	if (pxm_to_node_map[pxm] == NUMA_NO_NODE || node < pxm_to_node_map[pxm])
		pxm_to_node_map[pxm] = node;
	if (node_to_pxm_map[node] == PXM_INVAL || pxm < node_to_pxm_map[node])
		node_to_pxm_map[node] = pxm;
}

int acpi_map_pxm_to_node(int pxm)
{
	int node;

	if (pxm < 0 || pxm >= MAX_PXM_DOMAINS || numa_off)
		return NUMA_NO_NODE;

	node = pxm_to_node_map[pxm];

	if (node == NUMA_NO_NODE) {
		if (nodes_weight(nodes_found_map) >= MAX_NUMNODES)
			return NUMA_NO_NODE;
		node = first_unset_node(nodes_found_map);
		__acpi_map_pxm_to_node(pxm, node);
		node_set(node, nodes_found_map);
	}

	return node;
}

/**
 * acpi_map_pxm_to_online_node - Map proximity ID to online node
 * @pxm: ACPI proximity ID
 *
 * This is similar to acpi_map_pxm_to_node(), but always returns an online
 * node.  When the mapped node from a given proximity ID is offline, it
 * looks up the node distance table and returns the nearest online node.
 *
 * ACPI device drivers, which are called after the NUMA initialization has
 * completed in the kernel, can call this interface to obtain their device
 * NUMA topology from ACPI tables.  Such drivers do not have to deal with
 * offline nodes.  A node may be offline when a device proximity ID is
 * unique, SRAT memory entry does not exist, or NUMA is disabled, ex.
 * "numa=off" on x86.
 */
int acpi_map_pxm_to_online_node(int pxm)
{
	int node, min_node;

	node = acpi_map_pxm_to_node(pxm);

	if (node == NUMA_NO_NODE)
		node = 0;

	min_node = node;
	if (!node_online(node)) {
		int min_dist = INT_MAX, dist, n;

		for_each_online_node(n) {
			dist = node_distance(node, n);
			if (dist < min_dist) {
				min_dist = dist;
				min_node = n;
			}
		}
	}

	return min_node;
}
EXPORT_SYMBOL(acpi_map_pxm_to_online_node);

static void __init
acpi_table_print_srat_entry(struct acpi_subtable_header *header)
{
	switch (header->type) {
	case ACPI_SRAT_TYPE_CPU_AFFINITY:
		{
			struct acpi_srat_cpu_affinity *p =
			    (struct acpi_srat_cpu_affinity *)header;
			pr_debug("SRAT Processor (id[0x%02x] eid[0x%02x]) in proximity domain %d %s\n",
				 p->apic_id, p->local_sapic_eid,
				 p->proximity_domain_lo,
				 (p->flags & ACPI_SRAT_CPU_ENABLED) ?
				 "enabled" : "disabled");
		}
		break;

	case ACPI_SRAT_TYPE_MEMORY_AFFINITY:
		{
			struct acpi_srat_mem_affinity *p =
			    (struct acpi_srat_mem_affinity *)header;
			pr_debug("SRAT Memory (0x%llx length 0x%llx) in proximity domain %d %s%s%s\n",
				 (unsigned long long)p->base_address,
				 (unsigned long long)p->length,
				 p->proximity_domain,
				 (p->flags & ACPI_SRAT_MEM_ENABLED) ?
				 "enabled" : "disabled",
				 (p->flags & ACPI_SRAT_MEM_HOT_PLUGGABLE) ?
				 " hot-pluggable" : "",
				 (p->flags & ACPI_SRAT_MEM_NON_VOLATILE) ?
				 " non-volatile" : "");
		}
		break;

	case ACPI_SRAT_TYPE_X2APIC_CPU_AFFINITY:
		{
			struct acpi_srat_x2apic_cpu_affinity *p =
			    (struct acpi_srat_x2apic_cpu_affinity *)header;
			pr_debug("SRAT Processor (x2apicid[0x%08x]) in proximity domain %d %s\n",
				 p->apic_id,
				 p->proximity_domain,
				 (p->flags & ACPI_SRAT_CPU_ENABLED) ?
				 "enabled" : "disabled");
		}
		break;

	case ACPI_SRAT_TYPE_GICC_AFFINITY:
		{
			struct acpi_srat_gicc_affinity *p =
			    (struct acpi_srat_gicc_affinity *)header;
			pr_debug("SRAT Processor (acpi id[0x%04x]) in proximity domain %d %s\n",
				 p->acpi_processor_uid,
				 p->proximity_domain,
				 (p->flags & ACPI_SRAT_GICC_ENABLED) ?
				 "enabled" : "disabled");
		}
		break;

	default:
		pr_warn("Found unsupported SRAT entry (type = 0x%x)\n",
			header->type);
		break;
	}
}

/*
 * A lot of BIOS fill in 10 (= no distance) everywhere. This messes
 * up the NUMA heuristics which wants the local node to have a smaller
 * distance than the others.
 * Do some quick checks here and only use the SLIT if it passes.
 */
static int __init slit_valid(struct acpi_table_slit *slit)
{
	int i, j;
	int d = slit->locality_count;
	for (i = 0; i < d; i++) {
		for (j = 0; j < d; j++)  {
			u8 val = slit->entry[d*i + j];
			if (i == j) {
				if (val != LOCAL_DISTANCE)
					return 0;
			} else if (val <= LOCAL_DISTANCE)
				return 0;
		}
	}
	return 1;
}

void __init bad_srat(void)
{
	pr_err("SRAT: SRAT not used.\n");
	acpi_numa = -1;
}

int __init srat_disabled(void)
{
	return acpi_numa < 0;
}

#if defined(CONFIG_X86) || defined(CONFIG_ARM64)
/*
 * Callback for SLIT parsing.  pxm_to_node() returns NUMA_NO_NODE for
 * I/O localities since SRAT does not list them.  I/O localities are
 * not supported at this point.
 */
void __init acpi_numa_slit_init(struct acpi_table_slit *slit)
{
	int i, j;

	for (i = 0; i < slit->locality_count; i++) {
		const int from_node = pxm_to_node(i);

		if (from_node == NUMA_NO_NODE)
			continue;

		for (j = 0; j < slit->locality_count; j++) {
			const int to_node = pxm_to_node(j);

			if (to_node == NUMA_NO_NODE)
				continue;

			numa_set_distance(from_node, to_node,
				slit->entry[slit->locality_count * i + j]);
		}
	}
}

/*
 * Default callback for parsing of the Proximity Domain <-> Memory
 * Area mappings
 */
int __init
acpi_numa_memory_affinity_init(struct acpi_srat_mem_affinity *ma)
{
	u64 start, end;
	u32 hotpluggable;
	int node, pxm;

	if (srat_disabled())
		goto out_err;
	if (ma->header.length < sizeof(struct acpi_srat_mem_affinity)) {
		pr_err("SRAT: Unexpected header length: %d\n",
		       ma->header.length);
		goto out_err_bad_srat;
	}
	if ((ma->flags & ACPI_SRAT_MEM_ENABLED) == 0)
		goto out_err;
	hotpluggable = ma->flags & ACPI_SRAT_MEM_HOT_PLUGGABLE;
	if (hotpluggable && !IS_ENABLED(CONFIG_MEMORY_HOTPLUG))
		goto out_err;

	start = ma->base_address;
	end = start + ma->length;
	pxm = ma->proximity_domain;
	if (acpi_srat_revision <= 1)
		pxm &= 0xff;

	node = acpi_map_pxm_to_node(pxm);
	if (node == NUMA_NO_NODE || node >= MAX_NUMNODES) {
		pr_err("SRAT: Too many proximity domains.\n");
		goto out_err_bad_srat;
	}

	if (numa_add_memblk(node, start, end) < 0) {
		pr_err("SRAT: Failed to add memblk to node %u [mem %#010Lx-%#010Lx]\n",
		       node, (unsigned long long) start,
		       (unsigned long long) end - 1);
		goto out_err_bad_srat;
	}

	node_set(node, numa_nodes_parsed);

	pr_info("SRAT: Node %u PXM %u [mem %#010Lx-%#010Lx]%s%s\n",
		node, pxm,
		(unsigned long long) start, (unsigned long long) end - 1,
		hotpluggable ? " hotplug" : "",
		ma->flags & ACPI_SRAT_MEM_NON_VOLATILE ? " non-volatile" : "");

	/* Mark hotplug range in memblock. */
	if (hotpluggable && memblock_mark_hotplug(start, ma->length))
		pr_warn("SRAT: Failed to mark hotplug range [mem %#010Lx-%#010Lx] in memblock\n",
			(unsigned long long)start, (unsigned long long)end - 1);

	max_possible_pfn = max(max_possible_pfn, PFN_UP(end - 1));

	return 0;
out_err_bad_srat:
	bad_srat();
out_err:
	return -EINVAL;
}
#endif /* defined(CONFIG_X86) || defined (CONFIG_ARM64) */

static int __init acpi_parse_slit(struct acpi_table_header *table)
{
	struct acpi_table_slit *slit = (struct acpi_table_slit *)table;

	if (!slit_valid(slit)) {
		pr_info("SLIT table looks invalid. Not used.\n");
		return -EINVAL;
	}
	acpi_numa_slit_init(slit);

	return 0;
}

void __init __weak
acpi_numa_x2apic_affinity_init(struct acpi_srat_x2apic_cpu_affinity *pa)
{
	pr_warn("Found unsupported x2apic [0x%08x] SRAT entry\n", pa->apic_id);
}

static int __init
acpi_parse_x2apic_affinity(struct acpi_subtable_header *header,
			   const unsigned long end)
{
	struct acpi_srat_x2apic_cpu_affinity *processor_affinity;

	processor_affinity = (struct acpi_srat_x2apic_cpu_affinity *)header;
	if (!processor_affinity)
		return -EINVAL;

	acpi_table_print_srat_entry(header);

	/* let architecture-dependent part to do it */
	acpi_numa_x2apic_affinity_init(processor_affinity);

	return 0;
}

static int __init
acpi_parse_processor_affinity(struct acpi_subtable_header *header,
			      const unsigned long end)
{
	struct acpi_srat_cpu_affinity *processor_affinity;

	processor_affinity = (struct acpi_srat_cpu_affinity *)header;
	if (!processor_affinity)
		return -EINVAL;

	acpi_table_print_srat_entry(header);

	/* let architecture-dependent part to do it */
	acpi_numa_processor_affinity_init(processor_affinity);

	return 0;
}

static int __init
acpi_parse_gicc_affinity(struct acpi_subtable_header *header,
			 const unsigned long end)
{
	struct acpi_srat_gicc_affinity *processor_affinity;

	processor_affinity = (struct acpi_srat_gicc_affinity *)header;
	if (!processor_affinity)
		return -EINVAL;

	acpi_table_print_srat_entry(header);

	/* let architecture-dependent part to do it */
	acpi_numa_gicc_affinity_init(processor_affinity);

	return 0;
}

static int __initdata parsed_numa_memblks;

static int __init
acpi_parse_memory_affinity(struct acpi_subtable_header * header,
			   const unsigned long end)
{
	struct acpi_srat_mem_affinity *memory_affinity;

	memory_affinity = (struct acpi_srat_mem_affinity *)header;
	if (!memory_affinity)
		return -EINVAL;

	acpi_table_print_srat_entry(header);

	/* let architecture-dependent part to do it */
	if (!acpi_numa_memory_affinity_init(memory_affinity))
		parsed_numa_memblks++;
	return 0;
}

static int __init acpi_parse_srat(struct acpi_table_header *table)
{
	struct acpi_table_srat *srat = (struct acpi_table_srat *)table;

	acpi_srat_revision = srat->header.revision;

	/* Real work done in acpi_table_parse_srat below. */

	return 0;
}

static int __init
acpi_table_parse_srat(enum acpi_srat_type id,
		      acpi_tbl_entry_handler handler, unsigned int max_entries)
{
	return acpi_table_parse_entries(ACPI_SIG_SRAT,
					    sizeof(struct acpi_table_srat), id,
					    handler, max_entries);
}

int __init acpi_numa_init(void)
{
	int cnt = 0;

	if (acpi_disabled)
		return -EINVAL;

	/*
	 * Should not limit number with cpu num that is from NR_CPUS or nr_cpus=
	 * SRAT cpu entries could have different order with that in MADT.
	 * So go over all cpu entries in SRAT to get apicid to node mapping.
	 */

	/* SRAT: System Resource Affinity Table */
	if (!acpi_table_parse(ACPI_SIG_SRAT, acpi_parse_srat)) {
		struct acpi_subtable_proc srat_proc[3];

		memset(srat_proc, 0, sizeof(srat_proc));
		srat_proc[0].id = ACPI_SRAT_TYPE_CPU_AFFINITY;
		srat_proc[0].handler = acpi_parse_processor_affinity;
		srat_proc[1].id = ACPI_SRAT_TYPE_X2APIC_CPU_AFFINITY;
		srat_proc[1].handler = acpi_parse_x2apic_affinity;
		srat_proc[2].id = ACPI_SRAT_TYPE_GICC_AFFINITY;
		srat_proc[2].handler = acpi_parse_gicc_affinity;

		acpi_table_parse_entries_array(ACPI_SIG_SRAT,
					sizeof(struct acpi_table_srat),
					srat_proc, ARRAY_SIZE(srat_proc), 0);

		cnt = acpi_table_parse_srat(ACPI_SRAT_TYPE_MEMORY_AFFINITY,
					    acpi_parse_memory_affinity, 0);
	}

	/* SLIT: System Locality Information Table */
	acpi_table_parse(ACPI_SIG_SLIT, acpi_parse_slit);

	if (cnt < 0)
		return cnt;
	else if (!parsed_numa_memblks)
		return -ENOENT;
	return 0;
}

static int acpi_get_pxm(acpi_handle h)
{
	unsigned long long pxm;
	acpi_status status;
	acpi_handle handle;
	acpi_handle phandle = h;

	do {
		handle = phandle;
		status = acpi_evaluate_integer(handle, "_PXM", NULL, &pxm);
		if (ACPI_SUCCESS(status))
			return pxm;
		status = acpi_get_parent(handle, &phandle);
	} while (ACPI_SUCCESS(status));
	return -1;
}

int acpi_get_node(acpi_handle handle)
{
	int pxm;

	pxm = acpi_get_pxm(handle);

	return acpi_map_pxm_to_node(pxm);
}
EXPORT_SYMBOL(acpi_get_node);
Commit	Line	Data
1da177e4 LT	1	/*
	2	* acpi_numa.c - ACPI NUMA support
	3	*
	4	* Copyright (C) 2002 Takayoshi Kochi <t-kochi@bq.jp.nec.com>
	5	*
	6	* ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
	7	*
	8	* This program is free software; you can redistribute it and/or modify
	9	* it under the terms of the GNU General Public License as published by
	10	* the Free Software Foundation; either version 2 of the License, or
	11	* (at your option) any later version.
	12	*
	13	* This program is distributed in the hope that it will be useful,
	14	* but WITHOUT ANY WARRANTY; without even the implied warranty of
	15	* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
	16	* GNU General Public License for more details.
	17	*
1da177e4 LT	18	* ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
	19	*
	20	*/
ac906a6d HG	21
	22	#define pr_fmt(fmt) "ACPI: " fmt
	23
1da177e4	24	#include <linux/module.h>
1da177e4 LT	25	#include <linux/init.h>
	26	#include <linux/kernel.h>
	27	#include <linux/types.h>
	28	#include <linux/errno.h>
	29	#include <linux/acpi.h>
3770442e	30	#include <linux/memblock.h>
b552a8c5	31	#include <linux/numa.h>
99759869 TK	32	#include <linux/nodemask.h>
99759869 TK	33	#include <linux/topology.h>
1da177e4	34
762834e8	35	static nodemask_t nodes_found_map = NODE_MASK_NONE;
762834e8 YG	36
762834e8 YG	37	/* maps to convert between proximity domain and logical node ID */
ffada891	38	static int pxm_to_node_map[MAX_PXM_DOMAINS]
b552a8c5	39	= { [0 ... MAX_PXM_DOMAINS - 1] = NUMA_NO_NODE };
ffada891	40	static int node_to_pxm_map[MAX_NUMNODES]
b552a8c5	41	= { [0 ... MAX_NUMNODES - 1] = PXM_INVAL };
762834e8	42
8df0eb7c	43	unsigned char acpi_srat_revision __initdata;
e84025e2	44	int acpi_numa __initdata;
8df0eb7c	45
f363d16f	46	int pxm_to_node(int pxm)
762834e8 YG	47	{
762834e8 YG	48	if (pxm < 0)
b552a8c5	49	return NUMA_NO_NODE;
762834e8 YG	50	return pxm_to_node_map[pxm];
	51	}
	52
f363d16f	53	int node_to_pxm(int node)
762834e8 YG	54	{
	55	if (node < 0)
	56	return PXM_INVAL;
	57	return node_to_pxm_map[node];
	58	}
	59
d79ed248	60	static void __acpi_map_pxm_to_node(int pxm, int node)
3484d798	61	{
0f9b75ef DR	62	if (pxm_to_node_map[pxm] == NUMA_NO_NODE \|\| node < pxm_to_node_map[pxm])
	63	pxm_to_node_map[pxm] = node;
	64	if (node_to_pxm_map[node] == PXM_INVAL \|\| pxm < node_to_pxm_map[node])
	65	node_to_pxm_map[node] = pxm;
3484d798 DR	66	}
3484d798 DR	67
8ff6f48d	68	int acpi_map_pxm_to_node(int pxm)
762834e8	69	{
99759869 TK	70	int node;
99759869 TK	71
aec03f89	72	if (pxm < 0 \|\| pxm >= MAX_PXM_DOMAINS \|\| numa_off)
99759869 TK	73	return NUMA_NO_NODE;
	74
	75	node = pxm_to_node_map[pxm];
762834e8	76
1bb25df0	77	if (node == NUMA_NO_NODE) {
762834e8	78	if (nodes_weight(nodes_found_map) >= MAX_NUMNODES)
b552a8c5	79	return NUMA_NO_NODE;
762834e8	80	node = first_unset_node(nodes_found_map);
3484d798	81	__acpi_map_pxm_to_node(pxm, node);
762834e8 YG	82	node_set(node, nodes_found_map);
	83	}
	84
	85	return node;
	86	}
	87
99759869 TK	88	/**
	89	* acpi_map_pxm_to_online_node - Map proximity ID to online node
	90	* @pxm: ACPI proximity ID
	91	*
	92	* This is similar to acpi_map_pxm_to_node(), but always returns an online
	93	* node. When the mapped node from a given proximity ID is offline, it
	94	* looks up the node distance table and returns the nearest online node.
	95	*
	96	* ACPI device drivers, which are called after the NUMA initialization has
	97	* completed in the kernel, can call this interface to obtain their device
	98	* NUMA topology from ACPI tables. Such drivers do not have to deal with
	99	* offline nodes. A node may be offline when a device proximity ID is
	100	* unique, SRAT memory entry does not exist, or NUMA is disabled, ex.
	101	* "numa=off" on x86.
	102	*/
	103	int acpi_map_pxm_to_online_node(int pxm)
	104	{
dc9e0a93	105	int node, min_node;
99759869 TK	106
	107	node = acpi_map_pxm_to_node(pxm);
	108
	109	if (node == NUMA_NO_NODE)
	110	node = 0;
	111
dc9e0a93	112	min_node = node;
99759869	113	if (!node_online(node)) {
dc9e0a93 DW	114	int min_dist = INT_MAX, dist, n;
dc9e0a93 DW	115
99759869 TK	116	for_each_online_node(n) {
	117	dist = node_distance(node, n);
	118	if (dist < min_dist) {
	119	min_dist = dist;
dc9e0a93	120	min_node = n;
99759869 TK	121	}
	122	}
	123	}
	124
dc9e0a93	125	return min_node;
99759869 TK	126	}
	127	EXPORT_SYMBOL(acpi_map_pxm_to_online_node);
	128
ae2c6dcf DR	129	static void __init
ae2c6dcf DR	130	acpi_table_print_srat_entry(struct acpi_subtable_header *header)
1da177e4	131	{
1da177e4	132	switch (header->type) {
15a58ed1	133	case ACPI_SRAT_TYPE_CPU_AFFINITY:
4be44fcd	134	{
15a58ed1 AS	135	struct acpi_srat_cpu_affinity *p =
15a58ed1 AS	136	(struct acpi_srat_cpu_affinity *)header;
3dda4481 HG	137	pr_debug("SRAT Processor (id[0x%02x] eid[0x%02x]) in proximity domain %d %s\n",
	138	p->apic_id, p->local_sapic_eid,
	139	p->proximity_domain_lo,
	140	(p->flags & ACPI_SRAT_CPU_ENABLED) ?
	141	"enabled" : "disabled");
4be44fcd	142	}
1da177e4 LT	143	break;
1da177e4 LT	144
15a58ed1	145	case ACPI_SRAT_TYPE_MEMORY_AFFINITY:
4be44fcd	146	{
15a58ed1 AS	147	struct acpi_srat_mem_affinity *p =
15a58ed1 AS	148	(struct acpi_srat_mem_affinity *)header;
b9ced18a CF	149	pr_debug("SRAT Memory (0x%llx length 0x%llx) in proximity domain %d %s%s%s\n",
	150	(unsigned long long)p->base_address,
	151	(unsigned long long)p->length,
3dda4481 HG	152	p->proximity_domain,
	153	(p->flags & ACPI_SRAT_MEM_ENABLED) ?
	154	"enabled" : "disabled",
	155	(p->flags & ACPI_SRAT_MEM_HOT_PLUGGABLE) ?
	156	" hot-pluggable" : "",
	157	(p->flags & ACPI_SRAT_MEM_NON_VOLATILE) ?
	158	" non-volatile" : "");
4be44fcd	159	}
1da177e4 LT	160	break;
1da177e4 LT	161
7237d3de	162	case ACPI_SRAT_TYPE_X2APIC_CPU_AFFINITY:
7237d3de SS	163	{
	164	struct acpi_srat_x2apic_cpu_affinity *p =
	165	(struct acpi_srat_x2apic_cpu_affinity *)header;
3dda4481 HG	166	pr_debug("SRAT Processor (x2apicid[0x%08x]) in proximity domain %d %s\n",
	167	p->apic_id,
	168	p->proximity_domain,
	169	(p->flags & ACPI_SRAT_CPU_ENABLED) ?
	170	"enabled" : "disabled");
7237d3de	171	}
7237d3de	172	break;
3dda4481	173
4bac6fa7 HG	174	case ACPI_SRAT_TYPE_GICC_AFFINITY:
	175	{
	176	struct acpi_srat_gicc_affinity *p =
	177	(struct acpi_srat_gicc_affinity *)header;
	178	pr_debug("SRAT Processor (acpi id[0x%04x]) in proximity domain %d %s\n",
	179	p->acpi_processor_uid,
	180	p->proximity_domain,
	181	(p->flags & ACPI_SRAT_GICC_ENABLED) ?
	182	"enabled" : "disabled");
	183	}
	184	break;
	185
1da177e4	186	default:
ac906a6d HG	187	pr_warn("Found unsupported SRAT entry (type = 0x%x)\n",
ac906a6d HG	188	header->type);
1da177e4 LT	189	break;
	190	}
	191	}
	192
39b8931b FY	193	/*
	194	* A lot of BIOS fill in 10 (= no distance) everywhere. This messes
	195	* up the NUMA heuristics which wants the local node to have a smaller
	196	* distance than the others.
	197	* Do some quick checks here and only use the SLIT if it passes.
	198	*/
40e31856	199	static int __init slit_valid(struct acpi_table_slit *slit)
39b8931b FY	200	{
	201	int i, j;
	202	int d = slit->locality_count;
	203	for (i = 0; i < d; i++) {
	204	for (j = 0; j < d; j++) {
	205	u8 val = slit->entry[d*i + j];
	206	if (i == j) {
	207	if (val != LOCAL_DISTANCE)
	208	return 0;
	209	} else if (val <= LOCAL_DISTANCE)
	210	return 0;
	211	}
	212	}
	213	return 1;
	214	}
	215
e84025e2 DD	216	void __init bad_srat(void)
	217	{
	218	pr_err("SRAT: SRAT not used.\n");
	219	acpi_numa = -1;
	220	}
	221
	222	int __init srat_disabled(void)
	223	{
	224	return acpi_numa < 0;
	225	}
	226
6525afdf HG	227	#if defined(CONFIG_X86) \|\| defined(CONFIG_ARM64)
	228	/*
	229	* Callback for SLIT parsing. pxm_to_node() returns NUMA_NO_NODE for
	230	* I/O localities since SRAT does not list them. I/O localities are
	231	* not supported at this point.
	232	*/
	233	void __init acpi_numa_slit_init(struct acpi_table_slit *slit)
	234	{
	235	int i, j;
	236
	237	for (i = 0; i < slit->locality_count; i++) {
	238	const int from_node = pxm_to_node(i);
	239
	240	if (from_node == NUMA_NO_NODE)
	241	continue;
	242
	243	for (j = 0; j < slit->locality_count; j++) {
	244	const int to_node = pxm_to_node(j);
	245
	246	if (to_node == NUMA_NO_NODE)
	247	continue;
	248
	249	numa_set_distance(from_node, to_node,
	250	slit->entry[slit->locality_count * i + j]);
	251	}
	252	}
	253	}
3770442e HG	254
	255	/*
	256	* Default callback for parsing of the Proximity Domain <-> Memory
	257	* Area mappings
	258	*/
	259	int __init
	260	acpi_numa_memory_affinity_init(struct acpi_srat_mem_affinity *ma)
	261	{
	262	u64 start, end;
	263	u32 hotpluggable;
	264	int node, pxm;
	265
	266	if (srat_disabled())
	267	goto out_err;
e0af261a DD	268	if (ma->header.length < sizeof(struct acpi_srat_mem_affinity)) {
	269	pr_err("SRAT: Unexpected header length: %d\n",
	270	ma->header.length);
3770442e	271	goto out_err_bad_srat;
e0af261a	272	}
3770442e HG	273	if ((ma->flags & ACPI_SRAT_MEM_ENABLED) == 0)
	274	goto out_err;
	275	hotpluggable = ma->flags & ACPI_SRAT_MEM_HOT_PLUGGABLE;
	276	if (hotpluggable && !IS_ENABLED(CONFIG_MEMORY_HOTPLUG))
	277	goto out_err;
	278
	279	start = ma->base_address;
	280	end = start + ma->length;
	281	pxm = ma->proximity_domain;
	282	if (acpi_srat_revision <= 1)
	283	pxm &= 0xff;
	284
	285	node = acpi_map_pxm_to_node(pxm);
e0af261a DD	286	if (node == NUMA_NO_NODE \|\| node >= MAX_NUMNODES) {
e0af261a DD	287	pr_err("SRAT: Too many proximity domains.\n");
3770442e HG	288	goto out_err_bad_srat;
	289	}
	290
e0af261a DD	291	if (numa_add_memblk(node, start, end) < 0) {
	292	pr_err("SRAT: Failed to add memblk to node %u [mem %#010Lx-%#010Lx]\n",
	293	node, (unsigned long long) start,
	294	(unsigned long long) end - 1);
3770442e	295	goto out_err_bad_srat;
e0af261a	296	}
3770442e HG	297
	298	node_set(node, numa_nodes_parsed);
	299
	300	pr_info("SRAT: Node %u PXM %u [mem %#010Lx-%#010Lx]%s%s\n",
	301	node, pxm,
	302	(unsigned long long) start, (unsigned long long) end - 1,
	303	hotpluggable ? " hotplug" : "",
	304	ma->flags & ACPI_SRAT_MEM_NON_VOLATILE ? " non-volatile" : "");
	305
	306	/* Mark hotplug range in memblock. */
	307	if (hotpluggable && memblock_mark_hotplug(start, ma->length))
	308	pr_warn("SRAT: Failed to mark hotplug range [mem %#010Lx-%#010Lx] in memblock\n",
	309	(unsigned long long)start, (unsigned long long)end - 1);
	310
	311	max_possible_pfn = max(max_possible_pfn, PFN_UP(end - 1));
	312
	313	return 0;
	314	out_err_bad_srat:
	315	bad_srat();
	316	out_err:
	317	return -EINVAL;
	318	}
6525afdf HG	319	#endif /* defined(CONFIG_X86) \|\| defined (CONFIG_ARM64) */
6525afdf HG	320
15a58ed1	321	static int __init acpi_parse_slit(struct acpi_table_header *table)
1da177e4	322	{
2fad9308	323	struct acpi_table_slit slit = (struct acpi_table_slit )table;
1da177e4	324
39b8931b	325	if (!slit_valid(slit)) {
ac906a6d	326	pr_info("SLIT table looks invalid. Not used.\n");
39b8931b FY	327	return -EINVAL;
39b8931b FY	328	}
1da177e4 LT	329	acpi_numa_slit_init(slit);
	330
	331	return 0;
	332	}
	333
beffbe54	334	void __init __weak
7237d3de SS	335	acpi_numa_x2apic_affinity_init(struct acpi_srat_x2apic_cpu_affinity *pa)
7237d3de SS	336	{
ac906a6d	337	pr_warn("Found unsupported x2apic [0x%08x] SRAT entry\n", pa->apic_id);
7237d3de SS	338	}
7237d3de SS	339
7237d3de SS	340	static int __init
	341	acpi_parse_x2apic_affinity(struct acpi_subtable_header *header,
	342	const unsigned long end)
	343	{
	344	struct acpi_srat_x2apic_cpu_affinity *processor_affinity;
	345
	346	processor_affinity = (struct acpi_srat_x2apic_cpu_affinity *)header;
	347	if (!processor_affinity)
	348	return -EINVAL;
	349
	350	acpi_table_print_srat_entry(header);
	351
	352	/* let architecture-dependent part to do it */
	353	acpi_numa_x2apic_affinity_init(processor_affinity);
	354
	355	return 0;
	356	}
	357
1da177e4	358	static int __init
7237d3de	359	acpi_parse_processor_affinity(struct acpi_subtable_header *header,
4be44fcd	360	const unsigned long end)
1da177e4	361	{
15a58ed1	362	struct acpi_srat_cpu_affinity *processor_affinity;
1da177e4	363
15a58ed1	364	processor_affinity = (struct acpi_srat_cpu_affinity *)header;
1da177e4 LT	365	if (!processor_affinity)
	366	return -EINVAL;
	367
	368	acpi_table_print_srat_entry(header);
	369
	370	/* let architecture-dependent part to do it */
	371	acpi_numa_processor_affinity_init(processor_affinity);
	372
	373	return 0;
	374	}
	375
4bac6fa7 HG	376	static int __init
	377	acpi_parse_gicc_affinity(struct acpi_subtable_header *header,
	378	const unsigned long end)
	379	{
	380	struct acpi_srat_gicc_affinity *processor_affinity;
	381
	382	processor_affinity = (struct acpi_srat_gicc_affinity *)header;
	383	if (!processor_affinity)
	384	return -EINVAL;
	385
	386	acpi_table_print_srat_entry(header);
	387
	388	/* let architecture-dependent part to do it */
	389	acpi_numa_gicc_affinity_init(processor_affinity);
	390
	391	return 0;
	392	}
	393
095adbb6 TR	394	static int __initdata parsed_numa_memblks;
095adbb6 TR	395
1da177e4	396	static int __init
15a58ed1	397	acpi_parse_memory_affinity(struct acpi_subtable_header * header,
4be44fcd	398	const unsigned long end)
1da177e4	399	{
15a58ed1	400	struct acpi_srat_mem_affinity *memory_affinity;
1da177e4	401
15a58ed1	402	memory_affinity = (struct acpi_srat_mem_affinity *)header;
1da177e4 LT	403	if (!memory_affinity)
	404	return -EINVAL;
	405
	406	acpi_table_print_srat_entry(header);
	407
	408	/* let architecture-dependent part to do it */
095adbb6 TR	409	if (!acpi_numa_memory_affinity_init(memory_affinity))
095adbb6 TR	410	parsed_numa_memblks++;
1da177e4 LT	411	return 0;
	412	}
	413
15a58ed1	414	static int __init acpi_parse_srat(struct acpi_table_header *table)
1da177e4	415	{
2fad9308	416	struct acpi_table_srat srat = (struct acpi_table_srat )table;
1da177e4	417
8df0eb7c KG	418	acpi_srat_revision = srat->header.revision;
8df0eb7c KG	419
cfa806f0	420	/* Real work done in acpi_table_parse_srat below. */
1da177e4 LT	421
	422	return 0;
	423	}
	424
ae2c6dcf	425	static int __init
15a58ed1	426	acpi_table_parse_srat(enum acpi_srat_type id,
b43e1065	427	acpi_tbl_entry_handler handler, unsigned int max_entries)
1da177e4	428	{
6eb87fed	429	return acpi_table_parse_entries(ACPI_SIG_SRAT,
4be44fcd LB	430	sizeof(struct acpi_table_srat), id,
4be44fcd LB	431	handler, max_entries);
1da177e4 LT	432	}
1da177e4 LT	433
20e6926d	434	int __init acpi_numa_init(void)
e8d19552	435	{
20e6926d YL	436	int cnt = 0;
20e6926d YL	437
4bac6fa7 HG	438	if (acpi_disabled)
	439	return -EINVAL;
	440
d3bd0588 YL	441	/*
	442	* Should not limit number with cpu num that is from NR_CPUS or nr_cpus=
	443	* SRAT cpu entries could have different order with that in MADT.
	444	* So go over all cpu entries in SRAT to get apicid to node mapping.
	445	*/
8716273c	446
02cb489b	447	/* SRAT: System Resource Affinity Table */
7f8f97c3	448	if (!acpi_table_parse(ACPI_SIG_SRAT, acpi_parse_srat)) {
4bac6fa7	449	struct acpi_subtable_proc srat_proc[3];
702b07fc LA	450
	451	memset(srat_proc, 0, sizeof(srat_proc));
	452	srat_proc[0].id = ACPI_SRAT_TYPE_CPU_AFFINITY;
	453	srat_proc[0].handler = acpi_parse_processor_affinity;
	454	srat_proc[1].id = ACPI_SRAT_TYPE_X2APIC_CPU_AFFINITY;
	455	srat_proc[1].handler = acpi_parse_x2apic_affinity;
4bac6fa7 HG	456	srat_proc[2].id = ACPI_SRAT_TYPE_GICC_AFFINITY;
4bac6fa7 HG	457	srat_proc[2].handler = acpi_parse_gicc_affinity;
702b07fc LA	458
	459	acpi_table_parse_entries_array(ACPI_SIG_SRAT,
	460	sizeof(struct acpi_table_srat),
	461	srat_proc, ARRAY_SIZE(srat_proc), 0);
	462
20e6926d	463	cnt = acpi_table_parse_srat(ACPI_SRAT_TYPE_MEMORY_AFFINITY,
fd3e4543	464	acpi_parse_memory_affinity, 0);
1da177e4 LT	465	}
	466
	467	/* SLIT: System Locality Information Table */
7f8f97c3	468	acpi_table_parse(ACPI_SIG_SLIT, acpi_parse_slit);
1da177e4	469
20e6926d YL	470	if (cnt < 0)
20e6926d YL	471	return cnt;
095adbb6	472	else if (!parsed_numa_memblks)
f3946fb6	473	return -ENOENT;
940fed2e	474	return 0;
1da177e4 LT	475	}
1da177e4 LT	476
d79ed248	477	static int acpi_get_pxm(acpi_handle h)
1da177e4	478	{
27663c58	479	unsigned long long pxm;
1da177e4 LT	480	acpi_status status;
	481	acpi_handle handle;
	482	acpi_handle phandle = h;
	483
	484	do {
	485	handle = phandle;
	486	status = acpi_evaluate_integer(handle, "_PXM", NULL, &pxm);
	487	if (ACPI_SUCCESS(status))
50dd0969	488	return pxm;
1da177e4	489	status = acpi_get_parent(handle, &phandle);
4be44fcd	490	} while (ACPI_SUCCESS(status));
1da177e4 LT	491	return -1;
1da177e4 LT	492	}
1e3590e2	493
486c79b5	494	int acpi_get_node(acpi_handle handle)
1e3590e2	495	{
962fe9c9	496	int pxm;
1e3590e2	497
1e3590e2	498	pxm = acpi_get_pxm(handle);
1e3590e2	499
962fe9c9	500	return acpi_map_pxm_to_node(pxm);
1e3590e2 YG	501	}
1e3590e2 YG	502	EXPORT_SYMBOL(acpi_get_node);